1: \documentclass[prb,twocolumn,floats,graphicx,superscriptaddress]{revtex4}

2: %\documentclass[prb,preprint,superscriptaddress,graphicx]{revtex4}

3: \usepackage{graphicx}

4: \addtolength{\textheight}{0.5cm}

5: %\documentstyle[twocolumn]{article}

6: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

7: %TCIDATA{OutputFilter=LATEX.DLL}

8: %TCIDATA{LastRevised=Fri Jun 28 14:34:18 2002}

9: %TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}

10: %TCIDATA{CSTFile=revtex.cst}

11: 

12: \newcommand{\Real}{\mathop{\rm Re}\nolimits}

13: \newcommand{\sgn}{\mathop{\rm sgn}\nolimits}

14: \newcommand{\G}{\widehat G}

15: \newcommand{\eq}[1]{(\ref{#1})}

16: \newcommand{\beq}{\begin{equation}}

17: \newcommand{\eeq}{\end{equation}}

18: \newcommand{\bqn}{\begin{eqnarray}}

19: \newcommand{\eqn}{\end{eqnarray}}

20: \newcommand{\nn}{\nonumber \\}

21: 

22: %\topmargin=0cm

23: 

24: \begin{document} \title{Silent Phase Qubit Based on $d$-Wave

25: Josephson Junctions }

26: 

27: \author{M.H.S. Amin}

28: \affiliation{D--Wave Systems Inc., 320-1985 Broadway, Vancouver,

29: B.C., V6J 4Y3, Canada}

30: 

31: \author{A.Yu.~Smirnov}

32: \affiliation{D--Wave Systems Inc., 320-1985 Broadway, Vancouver,

33: B.C., V6J 4Y3, Canada}

34: 

35: \author{A.M.~Zagoskin}

36: \affiliation{D--Wave Systems Inc., 320-1985 Broadway, Vancouver,

37: B.C., V6J 4Y3, Canada} \affiliation{Physics and Astronomy Dept.,

38: The University of British Columbia, 6224 Agricultural Rd.,

39: Vancouver, B.C., V6T 1Z1, Canada}

40: 

41: \author{T.~Lindstr\"{o}m}

42: \affiliation{Microtechnology and Nanoscience, Quantum Device

43: Physics Laboratory, Chalmers University of Technology and

44: G\"oteborg University, SE-412 96 G\"oteborg, Sweden}

45: 

46: \author{S.~Charlebois}

47: \affiliation{Microtechnology and Nanoscience, Quantum Device

48: Physics Laboratory, Chalmers University of Technology and

49: G\"oteborg University, SE-412 96 G\"oteborg, Sweden}

50: 

51: \author{T.~Claeson}

52: \affiliation{Microtechnology and Nanoscience, Quantum Device

53: Physics Laboratory, Chalmers University of Technology and

54: G\"oteborg University, SE-412 96 G\"oteborg, Sweden}

55: 

56: \author{A.Ya.~Tzalenchuk}

57: \affiliation{National Physical Laboratory, Queens Road,

58: Teddington, Middlesex TW11 0LW, UK}

59: 

60: \begin{abstract}

61: 

62: We report on design and fabrication of a new type of flux qubit

63: that capitalizes on intrinsic properties of submicron YBCO grain

64: boundary junctions.  The operating point is protected from the

65: fluctuations of the external fields, already on the classical

66: level; the effects of external perturbations are absent, to the

67: second or third order, depending on the character of the

68: coupling. We propose an experiment to observe quantum tunneling

69: and Rabi oscillations in the qubit. Estimate of the decoherence

70: due to fluctuations of the external flux is presented.

71: 

72: \end{abstract}

73: 

74: \maketitle

75: %\date{\today}

76: 

77: %\vspace{5mm}

78: 

79: %\section{Introduction}

80: Over the last few years a series of experiments

81: \cite{expt,mooij,vion} provided a conclusive evidence of quantum

82: superposition in meso- and macroscopic superconductors. The task

83: at hand is scaling up of the system, with two goals in mind: to

84: probe how far quantum superposition can be pushed into

85: macroscopic world, and to develop an element base for quantum

86: computing, which only becomes viable on the scale 10--100 qubits.

87: 

88: It was suggested, that use of high-T$_c$ cuprates in qubits would

89: dispose of the need to apply fine-tuned external fields to keep

90: it in the operating point, due to the time-reversal symmetry

91: breaking in systems with DD junctions\cite{Blatter,zagoskin}. On

92: the level of a few qubits this is not a major advantage, compared

93: to the relative difficulty of fabrication and threat of extra

94: decoherence from nodal quasiparticles and zero-energy states (ZES)

95: in cuprates. Therefore the research was concentrated on

96: conventional superconductors, where all of the aforementioned

97: successes were achieved.

98: 

99: \begin{figure}[t]

100: \includegraphics[width=8cm]{fig1.ps}

101: \caption{{\protect\small (a) ``Silent" qubit. The dashed line

102: shows the location of the grain boundary junction. (b) Suggested

103: scheme for observation of Rabi oscillation and measurement of

104: decoherence. (c) AFM picture of a fabricated qubit with a

105: dc-SQUID fabricated around it for readout of the states.}}

106: \label{fig1}

107: \end{figure}

108: 

109: Nevertheless, the logical progression of research from single

110: qubit to qubit-qubit coupling and further to qubit registers is

111: bringing us to the point where intrinsic bistability of

112: high-T$_c$ qubits will become a major advantage. In the last few

113: years, we saw the development of reliable fabrication of bistable

114: submicron high-T$_c$ structures\cite{Ilichev01,Ivanov,Lindstrom}.

115: The danger of nodal quasiparticles and ZES seems to be

116: overestimated\cite{AminSmirnov,Fominov}. Finally, putting aside

117: any technological motivation, it is tempting to probe high-T$_c$

118: compounds for the same kind of macroscopic quantum coherence as

119: was observed in low-T$_c$ superconductors, just to see how and

120: whether these phenomena depend on, e.g., symmetry of the ground

121: state of the system.

122: 

123: The ``silent qubit" (for the etimology, see below) can be viewed

124: as a SQUID formed out of high-T$_c$ film, with two DD-grain

125: boundary junctions forming a mesoscopic island

126: (Fig.~\ref{fig1}a,c). The crystal lattice and $d$-wave order

127: parameter orientations on the two sides of the grain boundary are

128: chosen in such a way that each junction has doubly degenerate

129: ground state \cite{Ilichev01,amin}.

130: 

131: 

132: To simplify the analysis, we keep only two harmonics in the

133: current-phase relations of the DD junctions: $ I_{i}=I_{ci}\sin

134: \phi _{i}-I_{ci}^{\prime }\sin 2\phi _{i}$, where $I_{ci}$ is the

135: Josephson current, and $\phi _{i}$ is the phase difference across

136: the $i^{{\rm th}}$ junction ($i=1,2$). This approximation

137: successfully reproduces the observed behaviour of similar devices

138: in classical regime\cite{Ilichev01,Lindstrom}. The flux

139: quantization condition binds the phases $\phi _{1}$ and $\phi

140: _{2}$   to the total flux through the loop $\Phi$ via $\phi

141: \equiv \phi_{1}+\phi_{2}=2\pi \Phi /\Phi_{0}$, where $\Phi

142: _{0}=h/2e$ is the flux quantum. Introducing the superconducting

143: phase of the island $\theta \equiv (\phi _{1}-\phi _{2})/2$, the

144: free energy of the qubit, threaded by an external magnetic flux

145: $\Phi_x$,  in units of the Josephson energy $E_J$ ($\equiv E_1$,

146: where $E_{i}=I_{ci}\Phi_0/2\pi$), is given by

147: %

148: \begin{equation}

149: {\cal U}(\theta ,\phi )

150: %\equiv \frac{U_{J}}{E_{1}}

151: = {\frac{(\phi -\phi _{x})^{2}}{2{\beta_L}}} -{\cal E}_{\phi

152: }\left[ \cos \theta -\frac{\tilde{\alpha}_\phi}{4} \cos (2\theta

153: )\right] +\tilde{\cal U}(\theta ,\phi ). \label{Ut}

154: \end{equation}

155: Here   $\alpha _{i} = 2I_{ci}^{\prime}/I_{ci},$   ${\cal E}_{\phi

156: }=(1+\eta )\cos (\phi/2), \tilde{\alpha}_\phi= \left[(\alpha

157: _{1}+\eta \alpha _{2})\cos \phi\right]/\left[(1+\eta) \cos \left(

158: \phi /2\right)\right],$ $\eta =E_{2}/E_{1}{,}$ and $\phi _{x}=2\pi

159: \Phi _{x}/\Phi _{0}.$ The dimensionless self-inductance of the

160: loop ${\beta_L} \equiv 2\pi LI_{c1}/\Phi_0$ is considered

161: negligible (${\beta_L}\to 0$); then $\phi \to \phi _{x}$ and the

162: first term in (\ref{Ut}) can be dropped.

163: 

164: The last term in (\ref{Ut}):

165: $$

166: \tilde{\cal U}(\theta ,\phi )=-\left[ \eta -1+(\alpha _{1}-\eta

167: \alpha _{2})\cos \frac{\phi }{2}\cos \theta \right] \sin

168: {\frac{\phi }{2}}\sin \theta ,

169: $$

170: is zero when the two junctions are identical (i.e. $\eta =1$ and $%

171: \alpha _{1}=\alpha _{2}$). In this case the potential energy

172: minima are at $\theta = \pm \arccos (1/\tilde{\alpha}_\phi) $ when

173: $\tilde{\alpha}_\phi>1$ and zero otherwise. The $\pm$ signs

174: correspond to the states on the right and left sides of the

175: potential well (i.e. $|\pm\rangle$), respectively. For

176: $\tilde{\alpha}_{\phi}>1$ the potential has two minima, which are

177: degenerate at {\em any} external flux $\phi _{x}.$  The current

178: induced in the loop by the external flux does {\em not} depend on

179: the state of the qubit, which justifies the moniker ``silent". The

180: potential profile (\ref{Ut}) is similar to the one of persistent

181: current qubit \cite{mooij}, but here we have only one independent

182: phase (as ${\beta_L} \to 0$), and the problem becomes

183: one-dimensional. The barrier between the potential minima is

184: flux-dependent: $ W={\cos (\phi_x /2)}(

185: \tilde{\alpha}_\phi+\tilde{\alpha}_\phi^{-1}-2).$

186: 

187: 

188: In the general case ($\alpha_1\neq\alpha_2,$ $\eta\neq 1$) the

189: two minima are only degenerate when $\phi _{x}=0$. Now a {\em

190: state-dependent} persistent current flows in the loop even in

191: zero external field; in units of $I_{c1}$, ${\cal I}_{0}^{\pm

192: }=\pm [\eta \sqrt{\tilde{\alpha}_0^{2}-1}/(1+\eta

193: )\tilde{\alpha}_0^{2}](\alpha _{2}-\alpha _{1})$,

194: %

195: %\begin{equation}

196: %{\cal I}_{0}^{\pm }=\pm {\frac{\eta

197: %\sqrt{\tilde{\alpha}_0^{2}-1}}{(1+\eta

198: %)\tilde{\alpha}_0^{2}}}(\alpha _{2}-\alpha _{1})  \label{I}

199: %\end{equation}

200: where $\tilde{\alpha}_0 \equiv \tilde{\alpha}_{\phi =0}$.

201: 

202: An intermediate regime ($\alpha_1=\alpha_2$, $\eta\neq 1$) is most

203: interesting. It takes place when junctions have similar

204: current-phase dependences, but different critical currents (i.e.

205: widths), and should be expected if the junctions are close enough

206: to each other (Fig.~\ref{fig1}). At $\phi_x=0$, the equilibrium

207: value of $\theta$ is the same for both junctions, and there is no

208: spontaneous current: ${\cal I}_{0}^{\pm }=0$. At finite $\phi_x$,

209: the induced currents differ  for the two states of the qubit, but

210: the difference is of higher order in $\phi_x$, keeping the qubit

211: silent.

212: 

213: 

214: Expanding the free energy   to the third order in $\phi _{x}$, we

215: find for the minima:

216: %

217: \begin{equation}

218: {\cal U}_{{\rm min}}^{\pm }=A_{0}+A_{2}\phi _{x}^{2}\pm A_{3}\phi

219: _{x}^{3} \label{Umin}

220: \end{equation}

221: where $A_i$ are explicit, but cumbersome, functions of

222: $\tilde{\alpha}_0$ and $\eta$. As expected, there is no first

223: order dependence on $\phi_{x}$ ($A_{1}=0 $). The second order

224: term in (\ref{Umin}) does not depend on the state of the qubit.

225: The first state-dependent term in the minimum energy of the

226: system is $O(\phi _{x}^3)$ (Fig.~\ref{fig2}). Therefore small

227: fluctuations of $\phi _{x}$ do not affect the degeneracy of the

228: states. The difference between the energy minima grows as the

229: external flux is increased until the point at which the potential

230: barrier vanishes altogether, and the minimum with higher energy

231: disappears (the jumps in Fig.~\ref{fig2}).

232: 

233: %

234: \begin{figure}[t]

235: \includegraphics[width=8.7cm]{fig2.eps}

236: \caption{{\protect\small The minimum of the free energy (a) and

237: the corresponding values of $\theta$ at the minimum points (b) as

238: a function of $\phi_x$, for $\eta=2$ and $\alpha_1=\alpha_2=3$.

239: The solid and dashed lines correspond to the $\left|-\right>$ and

240: $\left|+\right>$ states respectively.}} \label{fig2}

241: \end{figure}

242: 

243: The current in the loop is found from $ {\cal I}_{\phi }^{\pm } =

244: \partial_{\phi_x} {\cal U}_{\rm min}^\pm = 2A_{2}\phi _{x}\pm

245: 3A_{3}\phi _{x}^{2}$. This current generates a state-dependent

246: magnetic flux $\delta \phi \equiv \phi -\phi _{x}= {\beta_L}{\cal

247: I}_{\phi }^{\pm }$. Note that the state-dependent contribution to

248: the induced flux is $O(\phi_x^2).$ For finite self-inductances

249: ${\beta_L}$, $\delta \phi$ has to be calculated

250: self-consistently. Figure~\ref{fig3} shows the result of such a

251: calculation. With the parameters chosen, an external flux close

252: to $0.2\Phi _{0}$ generates an additional flux ($\sim 0.005\Phi

253: _{0}$) through the loop, which is of the same order as the

254: estimate based on the above expansion.

255: 

256: Tunneling between the potential minima occurs due to the

257: uncertainty relation between the charge $Q$ of the island and its

258: superconducting phase $\theta$.  The tunneling matrix element is

259: approximately given by ($\hbar=k_{\rm B}=1$) $ \Delta \approx

260: \omega_p(\phi_x) e^{-\sqrt{\zeta W(\phi_x) {E_J/ E_c} }}$, where

261: $\zeta$ is a constant of the order of 1. The coefficient

262: $\omega_p(\phi_x) \equiv \sqrt{\omega_p^+\omega_p^-}$ is

263: determined by the frequencies $\omega_p^\pm$ of small oscillations

264: in the right and left potential minima, respectively. This

265: dependence follows from the expression of $\Delta$ as the matrix

266: element between the lowest energy states in the two wells. In the

267: case of $\tilde{\alpha}_0 \sim 1$, it is only valid

268: qualitatively\cite{TZA}, but enough for the present analysis. Due

269: to the symmetry of the potential profile when

270: $\alpha_1=\alpha_2$, the linear dependence on $\phi_x$ cancel and

271: we are left with $\omega_p(\phi_x) = \sqrt{E_J E_C (\alpha_0 -

272: \alpha_0^{-1})} (1-\kappa\phi_x^2), $ where $E_C = e^2 /2C$ is

273: the charging energy, $C$ is the effective capacitance of the

274: junctions, and $\kappa $ is a dimensionless coefficient of $O(1)$.

275: Fluctuations of $\phi_x$ influence $W$ and therefore $\Delta$.

276: Expanding the Josephson potential near the origin, we obtain the

277: tunnelling barrier $W={\cal U}_{\max }-{\cal U}_{\min }^{\pm

278: }=(B_{0}-A_{0})+\left( B_{2}-A_{2}\right) \phi _{x}^{2} +

279: O(\phi_x^3) $, where $B_{0} =-\left( \eta +1\right) \left(

280: 1-\tilde{\alpha}_0/4\right),$ $B_{2} = -\left[(\tilde{\alpha}_0

281: -1)/4\right]\left[\eta/(\eta +1)\right]$. Again there is no

282: dependence on $\phi_x$ in the lowest order.

283: 

284: Truncating the Hilbert space of the qubit to the two lowest energy

285: states, one can write the effective Hamiltonian of the system as $

286: H=\frac{1}{2}\Delta (\phi _{x})\sigma _{x}+\frac{1}{2}\epsilon

287: (\phi _{x})\sigma _{z}$, where $\epsilon (\phi _{x})\approx

288: E_{J}A_{3}\phi _{x}^{3}$. All single qubit operations can be

289: realized by applying controlled flux $\phi _{x}.$ Note that the

290: qubit only leaves the operating point when a finite external flux

291: is applied. Unlike the earlier qubit designs\cite{mooij,vion} this

292: point is protected from external flux fluctuations, already on

293: the classical level [cf. Eq.~(\ref{Umin})].

294: 

295: %

296: \begin{figure}[t]

297: \includegraphics[width=55mm]{fig3.ps}

298: \caption{\protect\small Self-consistent calculation of the

299: self-generated flux for the system of Fig.~\ref{fig2} with

300: ${\beta_L}=0.01$. } \label{fig3}

301: \end{figure}

302: 

303: Readout of the quantum state can in principle be achieved using a

304: dc-SQUID to measure the magnetic flux generated by the qubit when

305: it is biased by an external magnetic field. In this work the

306: readout SQUID was fabricated onchip using a high-$T_c$ grain

307: boundary SQUID (cf. Fig.~\ref{fig1}c); alternatively it can be

308: made of aluminum. Although coherent oscillations in an

309: all-aluminum persistent current qubit have previously been

310: observed using a similar readout design\cite{mooij}, the measured

311: decoherence time appeared to be rather short. Subsequent results

312: of Il'ichev et al.\cite{Rabi} have demonstrated a longer

313: decoherence time in a measurement scheme which couples the qubit

314: to a high quality tank circuit. Here, we calculate the decoherence

315: time of our silent qubit in the same setup\cite{Grajcar,Rabi}

316: (see Fig.~\ref{fig1}b). The circuit is assumed to have a

317: resonance frequency $\omega_T$ and the damping rate $\gamma_T$.

318: The mutual inductance is $M = k\sqrt{L_T L}$, $k \leq 1$ being

319: the coupling coefficient, and $L$ and $L_T$ are the inductance of

320: the qubit and tank respectively.

321: 

322: Without limiting the generality of our approach, we can ascribe

323: all the dephasing and dissipation in the qubit to its interaction

324: with the fluctuating flux $\phi_x(t)$, created by the tank current

325: $I_T(t)$. These fluctuations are characterized by correlator

326: $K(t,t')= \langle I_T(t), I_T(t') \rangle$, spectral density

327: $

328: K(\omega ) = \omega (\gamma_T \omega_T^2 / L_T) [\coth(\omega/2T) + 1] / [(\omega^2 - \omega_T^2)^2 +

329: \gamma_T^2 \omega^2 ],$

330: and dispersion $\langle I_T^2\rangle = K(t,t)$ $= ( \omega_T /

331: 2L_T) \coth (\omega_T /2T).$ At $\phi_x=0$, the qubit Hamiltonian

332: becomes $ H = (\Delta /2) \sigma_x  - \lambda_3 I_T^3

333: \sigma_z - \lambda_2 I_T^2 \sigma_x$, where $\lambda_3$ and

334: $\lambda_2$ are coupling coefficients depending on the qubit

335: parameters. In the Bloch-Redfield approximation\cite{smirnov}, we

336: calculate the energy relaxation rate $\Gamma = 30 \lambda_3^2

337: \langle I_T^2\rangle^2 (\gamma_T / \Delta^3) (\omega_T^2 /

338: L_T) \coth(\Delta /2T)$, together with a

339: dephasing time of the qubit $\gamma^{-1},$ where $ \gamma =

340: \Gamma/2 + \gamma_0 $,  $ \gamma_0 = (16 \pi / 3) \lambda_2^2

341: (\gamma_T^2 T^3 / L_T^2 \omega_T^4),$ if $T \ll \omega_T,$ and

342: $\gamma_0 = \lambda_2^2 Q_T (\omega_T / L_T^2) [

343: \coth^2(\omega_T / 2T) - 1 ] $ at

344: temperatures $T\geq \omega_T.$

345: 

346: Using the experimental data of Ref.~\onlinecite{TZA} ($I_c=0.5$

347: $\mu$A, $I_c'=0.6$ $\mu$A, $C\approx 10$ fF), we find $E_C \approx

348: 2$ GHz, $\omega_p/2\pi \approx 40$ GHz, and $\Delta/2\pi \approx

349: 1.6$ GHz. For ${\beta_L}\sim 0.01$, which is the value used in

350: Fig.~\ref{fig3}, the inductance of the loop is of the order of

351: $10pH$. To estimate the contributions of the cubic and quadratic

352: terms to qubit dephasing and dissipation, we chose the following

353: parameters: $\eta = 2,\ \alpha_1 = \alpha_2 = 2.4$, and $E_J

354: =1.66\times 10^{-22} J$.  If the tank frequency $\omega_T/2\pi =

355: 10$ MHz, its quality factor $Q_T = 2000,$ and the coupling

356: coefficient $ k \sim 1/33$, then contribution of {\em quadratic}

357: flux fluctuations to dephasing rate is small, so that the

358: dephasing time due to  qubit coupling to the tank is

359: $\gamma_0^{-1} \simeq 20$ ms at temperatures of order $10$ mK,

360: while the contribution of the {\em cubic} fluctuations to the

361: dephasing and relaxation rates is totally negligible. It means

362: that at the operating point the silent qubit is practically

363: decoupled from the fluctuations caused by the controlling

364: circuits. The dominant source of decoherence is from the nodal

365: quasiparticles at the junction, considered in

366: Ref.~\onlinecite{AminSmirnov}, which may reduce the decoherence

367: time to about 1--100 ns.

368: 

369: 

370: In conclusion, a new type of flux qubit, based on specific

371: properties of submicron YBCO grain boundary junctions, is proposed

372: and fabricated. The symmetry of the device provides an operating

373: point, which is stable and protected against the external field

374: fluctuations.

375: 

376: We are grateful to A. Golubov, E. Il'ichev, A. Maassen van den

377: Brink, Y. Nakamura, and A. Shnirman for stimulating discussions.

378: 

379: \begin{references}

380: %\begin{thebibliography}{99}

381: \bibitem{expt} Y. Nakamura, Yu. A. Pashkin, J.S. Tsai, Nature {\bf 398}, 786

382: (1999); J. Friedman et al., Nature {\bf 406}, 43 (2000); J.M.

383: Martinis, S. Nam, and J. Aumentado, Phys. Rev. Lett. {\bf 89},

384: 117901 (2002); Y. Yu, S. Han, X. Chu, and Z. Wang, Science {\bf

385: 296}, 889 (2002); A. J. Berkley {\it et al.}, Science {\bf 300},

386: 1548 (2003).

387: 

388: \bibitem{mooij}  I. Chiorescu {\it et al.},

389: Science {\bf 299}, 1869 (2003).

390: 

391: \bibitem{vion}  D. Vion {\it et al.}, Science {\bf 296}, 889 (2002).

392: 

393: \bibitem{Blatter} L.B. Ioffe {\it et al.},

394: %V.B. Geshkenbein, M.V. Feigel'man, A.L. Fauch\`{e}re, and G. Blatter,

395: Nature {\bf 398}, 679 (1999).

396: 

397: \bibitem{zagoskin}  A. Zagoskin, cond-mat/9903170; A. Blais and A.M.

398: Zagoskin, Phys. Rev. A {\bf 61}, 042308 (2000).

399: 

400: \bibitem{Ilichev01}  E. Il'ichev {\it et al.},

401: % M. Grajcar, R. Hlubina, R.P.J.IJsselsteijn, H.E. Hoenig, H.-G. Meyer, A. Golubov, M.H.S. Amin,

402: %A.M. Zagoskin, A.N. Omelyanchouk, and M.Yu. Kupriyanov,

403: Phys. Rev. Lett. {\bf 86}, 5369 (2001).

404: 

405: \bibitem{Ivanov} A.Ya. Tzalenchuk {\it et al.},

406: % T. Lindström, S. Charlebois, E.A. Stepantsov and Z. Ivanov,

407: ASC 2002 (Houston).

408: 

409: \bibitem{Lindstrom} T. Lindstr\"{o}m {\it et al.}, Phys. Rev. Lett. {\bf 90}, 117002 (2003).

410: 

411: \bibitem{AminSmirnov} M.H.S. Amin and A. Yu. Smirnov, preprint cond-mat/0304255.

412: 

413: \bibitem{Fominov} Ya. V. Fominov, A. A. Golubov, and M. Yu. Kupriyanov, JETP Lett.

414: {\bf 77}, 587 (2003).

415: 

416: \bibitem{amin}  M.H.S. Amin, A.N. Omelyanchouk, and A.M. Zagoskin, Phys.

417: Rev. B {\bf 63}, 212502 (2001); M.H.S. Amin {\it et al.},

418: % A.N. Omelyanchouk, S.N. Rashkeev, M. Coury, and A.M. Zagoskin,

419: Physica

420: B {\bf 318}, 162 (2002).

421: 

422: \bibitem{TZA} A.Ya. Tzalenchuk {\em et al.}, Phys. Rev. B {\bf 68}, 100501(R) (2003).

423: 

424: \bibitem{Grajcar} M. Grajcar {\em et al.}, preprint cond-mat/0303657.

425: 

426: \bibitem{Rabi} E. Il'ichev {\it et al.}, Phys. Rev. Lett. {\bf

427: 91}, 097906 (2003).

428: 

429: \bibitem{smirnov} A.Yu. Smirnov, preprint cond-mat/0306004

430: 

431: \end{references}

432: %\end{thebibliography}

433: 

434: %\begin{figure}[t]

435: %\includegraphics[width=12cm]{fig1.ps}

436: %\caption{(a) ``Silent" qubit. The dashed line shows the location

437: %of the grain boundary junction. (b) Suggested scheme for

438: %observation of Rabi oscillation and measurement of decoherence.

439: %(c) AFM picture of a fabricated qubit with a dc-SQUID fabricated

440: %around it for readout of the states.} \label{fig1}

441: %\end{figure}

442: 

443: %

444: %\begin{figure}[t]

445: %\includegraphics[width=12cm]{fig2.eps}

446: %\caption{The minimum of the free energy (a) and the corresponding

447: %values of $\theta$ at the minimum points (b) as a function of

448: %$\phi_x$, for $\eta=2$ and $\alpha_1=\alpha_2=3$. The solid and

449: %dashed lines correspond to the $\left|-\right>$ and

450: %$\left|+\right>$ states respectively. \vspace{10cm}} \label{fig2}

451: %\end{figure}

452: %

453: 

454: %\pagebreak

455: 

456: %\begin{figure}[t]

457: %\includegraphics[width=80mm]{fig3.ps}

458: %\caption{Self-consistent calculation of the self-generated flux

459: %for the system of Fig.~\ref{fig2}, with ${\beta_L}=0.01$. }

460: %\label{fig3}

461: %\end{figure}

462: 

463: \end{document}

464: